Publications

Yu, YR; Fang, SB; Zhuo, W; Han, JH (2024). A Fast and Easy Way to Produce a 1-Km All-Weather Land Surface Temperature Dataset for China Utilizing More Ground-Based Data. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 62, 5002016.

Abstract
Land surface temperature (Ts) is one of the important parameters of the Earth's surface, but the temporal and spatial incompleteness of Ts data has severely limited its application in many important fields, such as climate change, extreme weather, numerical weather prediction, and agro-meteorological disasters. The MODIS daily Ts data are relatively high temporal and spatial resolution data, but they also have a large amount of missing data due to the influence of clouds or other atmospheric conditions. Gap filling is currently the only means of obtaining complete high spatial and temporal resolution remote sensing Ts data. However, the current gap filling results either fail to guarantee the filling accuracy due to the difficulty of obtaining ground observation data or fail to generalize the gap filling method over a large area due to its inefficiency. In this study, we first obtained the daily mean Ts (dmTs) data under MODIS clear-sky condition using multiple linear regression based on Ts data from nearly 2400 meteorological observation stations and then proposed a new method to fill Ts in the cloudy-sky condition. The validation with in situ data showed that the precision of filled Ts in clear-sky condition indicates that its root-mean-square error (RMSE) is between 1.52-2.36 K, and in cloud-sky condition, its RMSE is 2.73 K. It is confirmed that the method has the advantages of efficiency, simplicity, and accuracy and is the most suitable method for filling Ts data at national and continental levels. The all-weather 1-km dmTs data reconstructed in this study are of great value on urban heat island intensity studying, air temperature generating, drought monitoring, and other associated applications.

DOI:
10.1109/TGRS.2024.3368707

ISSN:
1558-0644